Search

Mismatch repair (MMR) gene deficiency is rarely observed in gliomas, a constitutional defect is associated with tumorigenesis in Lynch syndrome, and an acquired defect is associated with hypermutation after temozolomide treatment. However, the meaning of MMR gene deficiency in gliomas is unclear. Two cases of MMR-deficient glioblastomas are reported, and mutational status of oncogenes was compared between primary and recurrent tumor samples in a glioblastoma patient with Lynch syndrome. Additionally, the characteristics of MMR-deficient glioblastomas were analyzed using public glioma datasets to determine the meaning of MMR deficiency in gliomas. Case 1 was a glioblastoma patient with Lynch syndrome, and treatment with pembrolizumab for the recurrent tumor was temporarily effective for a short period. Comparison of mutational changes between primary and recurrent tumor samples showed many additional mutated genes associated with multiple signaling pathways in the recurrent tumor. Tumor recurrence and chemoresistance could be associated with intratumoral heterogeneity and accelerated tumor progression due to defects of multiple signaling pathways. Case 2 was a glioblastoma patient with acquired MMR gene deficiency, and she died of rapid progression of bone marrow metastases. This rare clinical course was considered to be associated with gene expression changes and heterogeneity that resulted from MMR gene deficiency. Two cases of MMR gene-deficient glioblastomas were presented, and their genetic characteristics suggested that their clinical courses could be associated with MMR gene deficiency.

Machine learning models for automated magnetic resonance image segmentation may be useful in aiding glioma detection. However, the image differences among facilities cause performance degradation and impede detection. This study proposes a method to solve this issue. We used the data from the Multimodal Brain Tumor Image Segmentation Benchmark (BraTS) and the Japanese cohort (JC) datasets. Three models for tumor segmentation are developed. In our methodology, the BraTS and JC models are trained on the BraTS and JC datasets, respectively, whereas the fine-tuning models are developed from the BraTS model and fine-tuned using the JC dataset. Our results show that the Dice coefficient score of the JC model for the test portion of the JC dataset was 0.779 ± 0.137, whereas that of the BraTS model was lower (0.717 ± 0.207). The mean Dice coefficient score of the fine-tuning model was 0.769 ± 0.138. There was a significant difference between the BraTS and JC models (p &lt, 0.0001) and the BraTS and fine-tuning models (p = 0.002), however, no significant difference between the JC and fine-tuning models (p = 0.673). As our fine-tuning method requires fewer than 20 cases, this method is useful even in a facility where the number of glioma cases is small.

We previously reported that there was a subgroup of IDH-mutated astrocytomas harboring only 19q-loss showing oligodendroglioma-like morphology and significantly longer overall survival (OS) compared with 19q-intact astrocytomas. To further explore the biological characteristics of this possible subgroup and obtain insight into the mechanism of their clinical behavior, we compared gene expression pattern between five 19q-loss and five 19q-intact IDH-mutated astrocytomas by microarray analysis. Comparing expression level of each genes between 19q-loss and 19q-intact astrocytomas,136 up-regulated genes and 203 down-regulated genes were extracted. Gene expression patterns of 19q-loss astrocytomas were partially different from that of 19q-intact astrocytomas. More down-regulated genes distributed on 19q and 4p, and more up-regulated genes distributed on 4q. Multiple genes associated with stem cell maintenance were down-regulated in 19q-loss astrocytomas, and genes associated with glioma progression were differentially expressed. Comparing expression patterns among 19q-loss astrocytomas and other IDH-mutant glioma subgroups using TCGA datasets by t-SNE analysis revealed that expression pattern of 19q-loss astrocytomas did not shift to that of oligodendrogliomas with 1p/19q codeletion but were a subgroup in astrocytomas. These results indicated that 19q-loss in astrocytomas was an acquired event different from 1p/19q codeletion in oligodendrogliomas, and better prognosis morphological features in 19q-loss astrocytomas were derived from differentially expressed genes associated with stem cell maintenance and glioma progression.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has been fixed in the paper.

We report an extremely rare case of pial arteriovenous fistula (AVF) caused by trauma. A 61-year-old man suffered from brain contusion by a traffic accident. He was neurologically normal on admission. However, his headache gradually worsened, and partial seizures occurred thereafter. He presented with general tonic seizure 7 days after the head injury. Magnetic resonance imaging demonstrated the exacerbation of brain edema and an abnormal vein near the contusion. Subsequent angiography showed a pial AVF, which was considered to be responsible for the brain edema. After treatment of the AVF by direct surgery, the brain edema was ameliorated. We should take into consideration the formation of vascular disease in cases with unexpected worsening of edema after brain injury.

Low-grade gliomas often undergo malignant progression, and these transformations are a leading cause of death in patients with low-grade gliomas. However, the molecular mechanisms underlying malignant tumor progression are still not well understood. Recent evidence indicates that epigenetic deregulation is an important cause of gliomagenesis; therefore, we examined the impact of epigenetic changes during malignant progression of low-grade gliomas. Specifically, we used the Illumina Infinium Human Methylation 450K BeadChip to perform genome-wide DNA methylation analysis of 120 gliomas and four normal brains. This study sample included 25 matched-pairs of initial low-grade gliomas and recurrent tumors (temporal heterogeneity) and 20 of the 25 recurring tumors recurred as malignant progressions, and one matched-pair of newly emerging malignant lesions and pre-existing lesions (spatial heterogeneity). Analyses of methylation profiles demonstrated that most low-grade gliomas in our sample (43/51; 84%) had a CpG island methylator phenotype (G-CIMP). Remarkably, approximately 50% of secondary glioblastomas that had progressed from low-grade tumors with the G-CIMP status exhibited a characteristic partial demethylation of genomic DNA during malignant progression, but other recurrent gliomas showed no apparent change in DNA methylation pattern. Interestingly, we found that most loci that were demethylated during malignant progression were located outside of CpG islands. The information of histone modifications patterns in normal human astrocytes and embryonal stem cells also showed that the ratio of active marks at the site corresponding to DNA demethylated loci in G-CIMP-demethylated tumors was significantly lower; this finding indicated that most demethylated loci in G-CIMP-demethylated tumors were likely transcriptionally inactive. A small number of the genes that were upregulated and had demethylated CpG islands were associated with cell cycle-related pathway. In summary, we demonstrated that characteristic DNA demethylation occurred during malignant progression of a subset of low-grade gliomas. The mechanisms underlying and consequences of such DNA demethylation should be studied further.